Accepted Manuscript

Stage-specific distribution models can predict eel (Anguilla anguilla) occurrence

during settlement in coastal lagoons

C. Leone, M. Zucchetta, F. Capoccioni, M.F. Gravina, P. Franzoi, E. Ciccotti

PII: S0272-7714(16)30005-1
DOI: 10.1016/j.ecss.2016.01.005
Reference: YECSS 5003

To appear in: Estuarine, Coastal and Shelf Science

Received Date: 21 April 2015

Revised Date: 21 October 2015
Accepted Date: 3 January 2016

Please cite this article as: Leone, C., Zucchetta, M., Capoccioni, F., Gravina, M.F., Franzoi, P., Ciccotti,
E., Stage-specific distribution models can predict eel (Anguilla anguilla) occurrence during settlement in
coastal lagoons, Estuarine, Coastal and Shelf Science (2016), doi: 10.1016/j.ecss.2016.01.005.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Stage-specific distribution models can predict eel (Anguilla anguilla) occurrence during settlement in

coastal lagoons.

C. Leone1*, M. Zucchetta2, F. Capoccioni3, M.F. Gravina1, P. Franzoi2 and E. Ciccotti1

1
Dipartimento di Biologia, Università degli Studi di Roma “Tor Vergata”, Roma, Italy.
2
CEMAS - Center for Estuarine and Coastal Marine Sciences (CEMAS), Dipartimento di Scienze

PT
Ambientali Informatica e Statistica, University Ca' Foscari Venezia, Italy.
3
Consiglio per la Ricerca in Agricoltura e l'Analisi dell'Economia Agraria (CREA) - Centro di Ricerca per la

RI
Produzione delle Carni e il Miglioramento Genetico (PCM).

*Corresponding author: E-mail: chiara.leone@uniroma2.it; [add tel and fax number at proof]

SC
Abstract

U
Eel (Anguilla anguilla) is a catadromous fish species typical of Mediterranean coastal lagoons, that currently
AN
suffers from several anthropogenic and natural impacts. These are thought to be the cause of a stock-wide

decline that this panmictic species is facing, in inland and coastal waters of Europe and North Africa. The
M

decline affects both adult phases and recruitment, i.e. glass eel arrival to coastal waters and their ascent to

inland waters. Quantitative features of eel recruitment reflect a transoceanic global scale, but also depend on
D

local environmental conditions, the latter also affecting settlement dynamics in transitional waters. There is
TE

only little information on the dynamics of these two processes in coastal lagoons, notwithstanding the

paramount importance of both in sustaining local stocks abundance and their demographic structure for this
EP

typical but also economically important inhabitant of Mediterranean lagoons, habitats that constitute an

important share of the eel distribution area. The present study aims, therefore, to clarify space and time
C

dynamics of local scale recruitment and of settlement in a coastal lagoon in the Mediterranean area, also by
AC

setting up a specific methodological approach. For this purpose, data from field surveys in combination with

Species Distribution Models (SDMs) have been used in order to relate distribution of eel juvenile stages to

the environmental conditions within the lagoon. Specifically, models were calibrated to quantify the

relationship between presence of juvenile eel and the main environmental drivers, with the aim of identifying

potential habitats for eel settlement within the lagoon. Results gained by modelling suggest certain spatial

and temporal colonization patterns for the juvenile eel in the Fogliano lagoon, a typical Mediterranean

coastal lake. The modelling approach has therefore proved to be a useful tool for predicting habitats for eel
1
 ACCEPTED MANUSCRIPT
recruitment at the local scale and settlement, because adequate to catch the spatio-temporal dimensions of the

processes under study, in coastal lagoon habitats.

Keywords: Eel, Anguilla anguilla, settlement, coastal lagoon, GAM, Mediterranean region.

1. Introduction

Coastal lagoons and estuaries are among the aquatic habitats with the highest biological productivity (Day,

PT
1989; Perez-Ruzafa and Marcos, 2012) and provide suitable habitats for many fish species that dwell in these

transitional water bodies to feed and grow (Pihl et al., 2002; Franco et al., 2008). Among lagoon fishes,

RI
several are marine migrant species that use these habitats as nursery grounds (Minello et al., 2003; Burke et

SC
al., 2009; Franco et al., 2010). One of these is the eel, Anguilla anguilla, a typical inhabitant of

Mediterranean lagoons, which is also one of the most economically important (Perez-Ruzafa and Marcos,

U
2012; Capoccioni et al., 2014; Cataudella et al., 2014; Aalto et al., 2015). The eel uses Mediterranean
AN
lagoons as migratory pathways and/or feeding grounds to complete its life cycle, and these habitats constitute

a significant fraction (amounting to over 400 lagoons for a total surface area of approx. 580,000 ha) of the
M

overall continental habitat of the European eel (Aalto et al., 2015), This teleostean is a catadromous

panmictic species that migrates as an adult from European and Northern African inland and coastal waters to
D

the ocean, to finally spawn in the Sargasso Sea, and back to coastal waters and then continental growing
TE

habitats at the larval (leptocephalus) and juvenile (glass eel) stages (van Ginneken et al., 2005; Als et al.,

2011; Miller et al., 2014).

EP

The eel currently suffers from a number of problems related to many impacts of anthropogenic nature:
C

overfishing, habitat loss, pollution and climate change (Harrison et al., 2014) are all thought to cause the

stock-wide decline that this species is facing throughout its entire distribution area. The general concern for
AC

the global stock of eel is primarily due to a prolonged decline of its recruitment throughout the continent

(Moriarty and Dekker, 1997; ICES, 1999 and ICES reports thereafter). Eel recruitment, which depends

primarily, at the global scale, on the magnitude of the transoceanic larval migration, is a crucial process

because it sustains the many local stocks. These are distributed across the entire distribution area in several

water systems of varying typology, i.e. river catchments, channels, lakes, wetlands, coastal lagoons and near

shore marine environments. The features of this decline have been summarized by Harrison et al., 2014, who

2
 ACCEPTED MANUSCRIPT
also review behavioural and environmental processes, modulating juvenile eel recruitment patterns to the

catchment level. The success and entity of such a process depends in fact not only on global scale processes

(Knights, 2003; Kettle et al., 2006; Bonhommeau et al., 2008; Pecariz et al., 2014; Gutierrez-Estrada and

Pulido-Calvo, 2015), but also on local environment conditions related to the hydrological, morphological and

meteorological features of a specific site, which modulate space and time dynamics of the ascent (Gascuel et

PT
al.,1995, Arribas et al., 2012; Harrison et al., 2014; Trancart et al., 2014; Aranburu et al., 2015).

The importance of recruitment within the species life cycle explains the attention paid by researchers to the

RI
larval (Miller et al., 2014 for a review) and glass eel (Harrison et al., 2014 for a review) stages. Recruitment

SC
time and space dynamics at the local scale are well known for estuaries and rivers (Beaulaton and

Castelnaud, 2005; Harrison et al., 2014), in particular for many systems located on the Atlantic side of

U
Europe. Less information is available for the southernmost estuaries of the eel European distribution area
AN
(Arribas et al., 2012) and even less for recruitment and settlement of juvenile eel in coastal lagoons

ecosystems, notwithstanding the importance that these habitats play for the eel in the southern part of its
M

distribution area. To our knowledge only Crivelli et al. (2008), Lecomte- Finiger and Razouls (1981) and

Finiger (1976) have tried to determine which factors are involved in the glass eel migration to Mediterranean
D

coastal lagoons.
TE

Whatever the general mechanisms and the specific role of environmental cues, recruitment and effective

settlement are crucial when estimating eel production and hence breeder escapement and their contribution to
EP

the global spawning stock biomass (Aalto et al., 2015). Although overall recruitment may be very low today

(Dekker 2003a), settlement in lagoons has always been partially dependent on factors such as lagoon
C

morphology and hydrology and related to the effects of meteorological patterns on glass eel movement
AC

(Finiger 1976; Crivelli et al., 2008). Taken together, these result in an extremely variable recruitment level to

lagoons which in turn makes it very difficult to evaluate if recruitment level is the main limiting factor for eel

production in these habitats or simply one of several.

Against this background, this study aims to understand the space and time dynamics of recruitment and

settlement in a typical Mediterranean coastal lagoon, by setting up a specific methodological approach

allowing to seize the processes and to try to elucidate the correlations regulating them. For this purpose, data

3
 ACCEPTED MANUSCRIPT
from field surveys in combination with Species Distribution Models (SDMs) have been used in order to link

space and time dynamics to the environmental conditions within the lagoon. SDMs use quantitative methods

to infer environmental requirements of species from conditions at known occurrences (Guisan and

Zimmermann, 2000; Elith and Graham, 2009). SDMs are largely used for ecological applications in land

ecosystems, but they also have increasingly been used in aquatic environments (Robinson et al., 2011) to

PT
evaluate the nursery role of coastal areas (Stoner et al., 2001, 2007; Eastwood et al., 2003; Le Pape et al.,

2003; Vaz et al., 2008) and of transitional waters (e.g. Zucchetta et al., 2010).

RI
In this study, stage-specific SDMs have been developed on the basis of ad hoc fieldwork, considering glass

SC
eel, elver and young 0+ yellow eel. The models were applied using a set of predictor variables, which were

chosen from the many known environmental drivers of glass eel migration and young eel movements and

U
behaviour in order to highlight their role for juvenile eel colonization of coastal lagoons. The results, besides
AN
providing further insight into a scarcely known aspect of eel ecology, could be of use for setting up and

improving management strategies in coastal lagoons, also bearing in mind their potential role towards eel
M

stock recovery.

2. Materials and methods

D

2.1 Study area

TE

The study site (Figure 1) is the Fogliano lagoon (41° 24' 12° 54), since 1975 included within the Parco

Nazionale del Circeo. The surface of the lagoon is 404 ha, mean depth is 0.89 m (2.00 m maximum depth). It
EP

is oriented parallel to the coastal line and - due to reclamation interventions carried out between 1926 and

1935 - the lagoon has a rectangular shape with straight shorelines. On the western side, a tidal channel
C

(length 270 m, width 16 m) connects the lagoon to the sea, while a second channel (length 495 m, width 38
AC

m) on the southern side used to connect the lagoon to the mouth of a nearby river, Rio Martino. In the late

1980s, this connection was interrupted by a sluice to prevent water exchanges with the river, due to its poor

water quality. Similarly, other surface tributaries, mostly located in the north-western part of the lagoon,

were diverted in a channel (the Cicerchia channel, Figure 1) to avoid excessive organic enrichment to the

lagoon and subsequent eutrophication. As a consequence, freshwater inputs are currently drastically reduced,

apart random runoffs from the continental side of the lagoon where occasional overflows can occur from a

perimeter channel along the inner shore of the lagoon (Figure 1) or occasional river overflows of the Martino
4
 ACCEPTED MANUSCRIPT
river mouth at incoming tides (Figure 1). Several authors have observed a general increase of salinity values

over time (e.g.: Mariani, 2001: annual average salinity 36 (range: 30-47); Signorini et al., 2008: 37 (range:

30 - 45); Manzo, 2010: 39 (range 29 – 49)).

Maximum range of spring tide amounts to about 23 cm, with 17 cm of range during the semidiurnal tide. In

normal sea conditions, at each tide the lagoon receives 42.300 m3 from the sea, while during the ebb tide it

PT
returns 37.700 m3; on the other hand during coastal storms there is a continuous flow into the lagoon.

Essentially, water volumes in and out the lagoon show a marked variability of outflows with differences in

RI
volumes in relation to the changing hydrological conditions that are established between the lagoon and the

sea (Progetto Laghi Costieri, 1985).

SC
As for management, in this lagoon, a small-scale artisanal fishery was present, targeting euryhaline fish

species and eel as well. The fishery was stopped in 2007, which resulted in the protection of recruiting and

U
growing stages, and the implementation of silver eel escapement (Ciccotti, pers. obs), due to ceased fishing
AN
pressure and related mortality.

2.2 Sampling
M

Three 7-day long sampling campaigns were carried out in December 2012, February and April 2013, in
D

coincidence with new moon, aiming at catching juvenile eels at different moments of the ascending seasons,
TE

and eventually in different phases of the settlement process. Samplings were carried out at seven stations

(Figure 1, numbered anticlockwise) chosen in order to cover most of the environmental variability of the
EP

lagoon: two stations are near the main tidal channel (1,7), one close to the channel leading to Martino river

(3), two stations in the central area of the lagoon (2, 4), one the continental side of the lagoon (5) and one in
C

the northern part of the lagoon (6).

AC

Samplings were aimed at catching all possible stages of juvenile eel: unpigmented glass eel (GLE),

pigmented elver (ELV) and small 0+ juvenile eel up to 15 cm, conventionally called ‘bootlace’ (BTL).

Zompola et al. (2008) highlight the importance of taking into account the ontogenetic stage of juvenile eel

when studying recruitment dynamics. This study targeted all the juvenile stages, and individuals caught by

samplings were then assigned to those three stages, GLE being the unpigmented or pigmenting stage up to

stage VIAII according to Strubberg (1913) and adapted by Bertin (1956), ELV being the stage completely

5
 ACCEPTED MANUSCRIPT
pigmented, with activated swim bladder (Hickman, 1981) and functional digestive tract and BTL being the

small growing yellow eel up to 15 cm.

Silberschneider et al. (2001) proposed and validated the use of artificial habitat collection devices for

studying resettlement patterns in anguillid glass eels. Therefore, specific traps (T) are designed and built for

this study using PVC tubes (30.0 cm diameter and 50.0 cm long) closed at one end by 1.0 mm mesh. Traps

PT
were filled with bundles of olive leaves, to encourage colonization, as Silberschneider et al. (2001) suggest.

In each station glass eel fyke nets (GEfy, 1.6 m long, 4 chambers, 40cm wide, with a 1.5 mm mesh size and

RI
wings of 1.3 m) and eel fyke nets (Efy, 2.7 m long, 6 chambers, mouth diameter of 40 cm, wings of 2.5 m

and a mesh size 4.0 mm) were also installed, simultaneously at each sampling campaign. In order to reduce

SC
artificial appearance and odour, all sampling devices were kept under water in the lagoon for three weeks

prior to the beginning of sampling, as also suggested by Silberschneider et al. (2001).

U
Sampling gears in each station were inspected daily for the whole duration of each campaign, for a total of
AN
twenty-one fishing gears a day (three gears x seven stations) over the whole lagoon. All captured individuals

were anaesthetized with clove oil (Walsh and Pease 2002) and then counted, weighed and their length
M

measured. Assessment of pigmentation (according to the scale of Strubberg, 1913), swim bladder activation
D

(Hickman, 1981) and gut content, examined in vivo by means of a stereo-microscope to verify the onset of
TE

swimming behaviour and food intake, allowed assignment to one of the three considered stages. All eels

were released alive in a nearby channel.

EP

Environmental parameters (temperature (T), salinity (Sal), dissolved oxygen (DO) and Chlorophyll-a) were

recorded daily at each sampling station with a multiparameter sonde (OxyGuard Handy Gamma). In March
C

2013, no in-field monitoring of the presence of eel was performed, but an extra campaign was carried out to
AC

gather additional environmental data (15 sampling stations all around the lagoon) and thereby expanding the

predictor variable dataset to apply SDMs. Distances between each sampling station and the tidal channel (Δ

inlet) or any possible fresh water supply (Δ FW) were measured by Geographic Information System (GIS), in

order to assess sea and fresh water influence. Even if semidiurnal tidal range was extremely reduced, the tide

phase was taken into account, considering it as predictor for the variable describing the lag (expressed as

days) from the maximum of the new moon spring tide.

6
 ACCEPTED MANUSCRIPT
Given the importance of considering biotic variables in ecological studies on fish in coastal lagoons

ecosystems, attention was given to the possible role of juvenile eel feeding behaviour in influencing

settlement. Bardonnet and Riera (2005) have shown that glass eels keep a marine isotopic signature during

migration but pigmented eels present an isotopic signature typical of inland sources, possibly originated from

benthic macro-invertebrates or small fishes. Therefore, among the possible biotic variables the choice has

PT
fallen on the zoo-benthic community and in particular on the abundance of some macro-benthic invertebrate

taxa, in order to portray prey abundance.

RI
Two replicate sediment samples were collected at each station by a Van Veen grab (250 cm3). Each sample

was sieved in water with a 0.5 mm sieve and immediately frozen and stored at - 20° C for successive sorting.

SC
After defrosting, organisms were extracted from the sediment, identified to the lower taxonomic level (genus

or species), counted and measured. For the identification, the following literature was used: Fauvel (1923,

U
1927), Ruffo (1982; 1988), Gravina and Somaschini (1990), Sconfietti (2004) and Doneddu and Trainito
AN
(2005). Among all benthic species, only those that could be considered prey/food for the juvenile eel were

retained, based on literature information (Tesch, 2003), size range (<0.10 cm), abundance and expert
M

judgement.
D

2.3 Development of Species Distribution Models (SDMs)

TE

Relationships between the presence of GLE, ELV and BTL and the predictor variables were studied by

means of Generalized Additive Models (GAM) (Hastie and Tibshirani, 1990). Specifically, data of presence
EP

and of abundance were coupled per eel stage and sampling station. A matrix of presence/absence was used

for setting up the models, while abundance data were used to observe changes in time and space of eel
C

distribution. Only predictors supported by strong biological reasoning, based on a priori investigation of the
AC

scientific literature, as suggested by Burnham and Anderson (2002), were used to generate set models by

means of combinations of predictors, representing different hypotheses on how environmental conditions

affect eel distribution (Table 3). To reduce the risk of collinearity in the model only variables with a Pearson

correlation lower than 0.7 (a common inclusion threshold) have been considered (Table A.1). Overall, the

predictor variables retained for further analysis were: temperature (T), salinity (Sal), dissolved oxygen (DO),

days after new moon (Moon), distance from the tidal channel (∆ inlet), distance from the freshwater input (∆

FW) and abundances (n) of six benthic invertebrate species representing potential preys: Chironomidae
7
 ACCEPTED MANUSCRIPT
larvae (Chi), Monocorophium insidiosum (Min), Microdeutopus gryllotalpa (Mgr), Gammarus aequicauda

(Gae), Platynereis dumerilii (Pdu) and Nereiphylla rubiginosa (Nru).

The most suitable model was chosen based on the framework of an information-theoretical approach

(Burnham and Anderson, 2002) and corresponded to the most reasonable hypothesis (i.e. better supported by

the data). The approach to GAM as proposed by Wood (2006) was followed for fitting the models, using the

PT
‘mgcv’ library (Family: binomial; Link function: logit) for the R software packages (R Development Core

Team, 2014). For each response variable (presence/absence of each eel stage), model selection is based on

RI
the Akaike Information Criterion, corrected for small sample sizes (AICc), considering the model with the

SC
lowest AICc as the best model of the set. When it was not possible to identify a clear best model (differences

of AICc between the two best model smaller than 4), models were combined following a model averaging

U
approach, based on AICc-derived weights (Freckleton, 2011). Finally, the averaged model was delineated
AN
using the 95% confidence of models (95% of the summed weight, Burnham and Anderson, 2002). The

dataset was split by random sampling, in order to use two independent sets for model calibration and
M

evaluation. About 70% of the data (103 observations) were used for model building, while the remaining

30% (44 observations) were used to test the models for each of the three stages. Considering the importance
D

of using different performance estimations for the habitat model (Zurell et al., 2009), we used several
TE

statistics to infer the model’s predictive capability: sensitivity (conditional probability that a presence is

correctly classified); specificity (conditional probability that an absence is correctly classified); percent of
EP

correctly classified observations (PCC); Cohen's kappa, and the area under the receiver operator curve

(AUC) (Fielding and Bell, 1997). Thresholds for the first four statistics were defined following the
C

minimized difference threshold criterion (difference between sensitivity and specificity) (Jimenez-Valverde
AC

and Lobo, 2007). To obtain the spatial predictions for the whole lagoon basin, the fitted models (or the

ensembles of fitted models) were applied in a predictive way using raster layers as predictors. All predictive

variables were interpolated by using ordinary kriging (library automap; Hiemstra et al., 2008) to generate

daily maps, and models predictions were carried out in order to predict temporal and spatial presence of the

early life stages of Anguilla anguilla during the recruitment and settlement phases. Monthly maps for

December, February and April for the three stages were obtained averaging the daily prediction maps.

8
 ACCEPTED MANUSCRIPT
3. Results

3.1 Presence and abundance

Juvenile eel abundances recorded in the three sampling campaigns (December, February and April) are

summarized in Table 1 and shown in Figure 2. Information on temperature, salinity, oxygen and chlorophyll

are also reported in order to illustrate the spatial and temporal environmental variation in lagoon, in support

PT
of the distribution of eel life stage presence (Table 2). GLE (glass eel) was the most frequent and the most

abundant stage present in December. Its peak of abundance occurred in February, while only one presence

RI
was recorded in April. In both campaigns, GLE was present at the two stations closest to the tidal channel

(St. 1, 7) and at the connection with the adjoining river (St. 3). ELV (elver), on the other hand, were caught

SC
on the inner side of the lagoon (St. 2, 3, 4, 5) where they were present in all campaigns, predominantly in

February. BTL (bootlace), were most abundant in April when they were the dominant stage and were

U
distributed evenly throughout the entire lagoon (all stations).
AN
3.2 SDM development

It was not possible to identify a single best model for any stage, and the best set of models used in the
M

averaging approach differed for the three stages, delineating different relationships with environmental
D

conditions (Table 3). Models included for GLE are complex because of the high number of variables and
TE

they include many biotic (prey abundance) as well as abiotic variables. Models for ELV describe its

distribution in relation to salinity, and distance to sea-inlet and freshwater inputs. A higher uncertainty is
EP

associated with the selection of models for BTL, as testified by the larger number of models included in the

95% confidence set and by the fact that among the models none seems to have a stronger weight than the
C

others.
AC

The statistics evaluated on the validation dataset, suggest that fitted models proved to have a good to very

good prediction capability comparing predicted probability of presence with observations (AUC - GLE:

0.80; ELV: 0.70; BTL: 0.75). Models predictions could also be transformed into presence/absence

predictions yielding a high percentage of records correctly classified (PCC - GLE: 0.70; ELV: 0.73;

BTL:0.73) and maintaining a good balance between omission and commission errors for glass eel and

bootlace but not for elvers, for which the model seems to underestimate presence (sensitivity - GLE: 0.73;

9
 ACCEPTED MANUSCRIPT
ELV: 0.40; BTL: 0.80; specificity - GLE: 0.70; ELV: 0.78; BTL: 0.72; Kappa - GLE: 0.35; ELV: 0.11; BTL:

0.28).

3.3 SDM application

The application of the models allowed to predict the probability of presence of the three juvenile eel stages in

the Fogliano lagoon as a function of environmental variables used as predictors (Figure 3). GLE-probability

PT
of presence is higher in association with lower salinity conditions and with temperature values between 5-10

°C. The probability of ELV-presence can also be predicted by salinity, but not by temperature, whereas the

RI
probability of BTL-presence is associated with higher temperatures.

SC
A decrease of the probability of presence of GLE and ELV is associated with the distance from the tidal

channel. Distance from the freshwater inlets, on the other hand, does not seem to influence the probability of

U
presence of GLE, though it affects slightly the older stages in the opposite way: ELV are more likely to be
AN
present closer to the freshwater inputs, whereas BTL are more likely to be present far from freshwater

channels. A negative relationship with oxygen content is recorded for GLE, while this variable does not
M

seem to affect ELV and BTL distributions. Probably this is related to the fact that the oxygen concentration

never reached critical values (minimum concentration: 5.7 mgl-1) and ELV is only indirectly associated to
D

this variable, i.e. GLE prefer conditions found in areas with a relatively low oxygen content. The moon phase
TE

seems to be hardly relevant in affecting probability of presence of GLE and it is not-relevant for older stages.

Concerning to the role played by potential prey in influencing juvenile eel stages distribution in the lagoon
EP

and against time, no role is shown by any of the benthic species considered for ELV nor for BTL, contrary to
C

what expected. On the other hand, selected sets of models (Table 3) show an association of probability of

presence of GLE with some benthic species such as the Amphipods Monocorophium insidiosum (Min),
AC

Microdeutopus gryllotalpa (Mgr) and Gammarus aequicauda (Gae) and the Polychaetes Nereiphylla

rubiginosa (Nru), Platynereis dumerilii (Pdu) and Chironomids larvae (Chi) (Figure 3).

SDMs application allowed to predict and display the probability of presence of GLE, ELV and BTL in the

whole lagoon basin. The monthly maps (Figure 4) generalise the patterns described by the daily maps

(Figures A.1, A.2 and A.3). Predictions for March were obtained on the basis of environmental data collected

during the extra campaign across the lagoon (see Section 2.2).

10
 ACCEPTED MANUSCRIPT
In detail, monthly maps show how GLE peak (in terms of probability of presence) is likely to occur in

December and February (Figure 4), with a space pattern that see GLE propagating from the tidal channel to

the rest of the lagoon. In the case of ELV, the probability of presence is higher in February and March and

with a concentration in the southern area of the lagoon. For BTL, for which a very complex set of models has

been retained (Table 3), only temperature seems to play a dominant role in the period of presence (April).

PT
BTL presence is homogeneous in terms of distribution across the lagoon, but with a higher probability of

presence in the northern area, the most confined sector of the lagoon.

RI
4. Discussion

4.1 Methodological aspects

SC
This study has several notable different aspects. One is methodological and concerns the sampling scheme

U
that proved to be adequate to catch the space-temporal dimensions of the processes, i.e.coastal lagoon
AN
juvenile eel recruitment and settlement. The combined use of three different sampling gears, aimed at

capturing different eel juvenile stages in an adequate number of sampling stations and at opportune time
M

intervals i.e. time windows significantly different one from another in terms of environmental conditions and

representing different phases of the settlement process, allowed to observe the evolution of juvenile eel
D

requirements in space and in time. The proof of this assumption was an important prerequisite for our work
TE

and of particular importance for the purpose of this study, as the conditions observed in a survey define the

accuracy, precision and generality (i.e. transferability in time and space) of a species distribution model
EP

(Austin, 2002). Highly variable environments, such as estuaries and lagoons, are characterized by marked

temporal dynamics in species-habitat associations (Gibson, 1997; Able et al., 2005), which is also true,
C

though, for the association between specific habitats and definite juvenile eel stages. The complex sampling
AC

scheme was therefore necessary to adequately cover the environmental variability of the recruitment period

in a typical Mediterranean lagoon and to catch some characteristics crucial for fitting SDMs. First, the

description of the spatio-temporal dynamic of colonisation allowed splitting the problem of model

development into a multi-stage approach, which is the optimal solution for species experiencing sharp

ontogenetic shifts (Robinson et al., 2011). Second, even if the distribution models were developed using

presence/absence data, the adequateness/robustness of the collected information, related to the frequency of

presence (and synthesised by the number of collected specimens for each stage), seems to influence model
11
 ACCEPTED MANUSCRIPT
accuracy and reliability (higher for GLE and BTL). Indeed, the strong imbalance between presence and

absence (stronger for ELV than for the other stages) may influence the capability of an empirical model to

link the distribution to environmental conditions. This could be the possible explanation for why models

underestimated the presence of elvers and not the other stages.

4.2 Ecological aspects

PT
SDMs development and application allowed predicting juvenile eel presence inside the lagoon on the basis

of a set of predictor variables, chosen from existing knowledge about environmental drivers of glass eel

RI
migration and young eel movements and behaviour. If we assume that the predictors that have yielded good

SC
results are the variables that influence recruitment dynamics, then these parameters are those that affect

juvenile eel colonization in the Fogliano coastal lake, and more generally in coastal lagoons. The results of

U
this study show that it is possible to describe the role of different groups of variables influencing recruitment,
AN
and further the settlement of the various stages.

The physical variable that plays the most influential role is temperature. For glass eels, the optimal
M

temperature range was between 5 and 10 °C, which influenced glass eel movements on an intra-seasonal

basis. The lower value is in accordance with what reported by most authors, i.e. that temperature values
D

lower then 5-4 °C render glass eel inactive, while it has been reported that when the difference in
TE

temperature between sea and lagoon exceeds 5 or 6 °C, glass eel entrance into colder waters may be

prevented (Gandolfi et al., 1984; Jessop, 2003; Edeline et al., 2006). On the other hand, seasonal water
EP

temperatures influence movements of the younger eel stages (elver and bootlace stages) within the lagoon.

Most studies explain that warmer water temperature is positively correlated with active swimming, growth,
C

metabolism and pigmentation in both Atlantic and Pacific eel (Haro and Krueger 1988; White and Knights,
AC

1997; August and Hicks 2006; Edeline et al., 2006; Luers et al., 2011).

The overall environmental gradients in the lagoon are also important, although to different degrees for the

three stages, as outlined by the role of chemical (salinity and to a lesser extent oxygen) and morphological

(distance from the tidal channel, distance from freshwater inputs) parameters. Within a lagoon,

environmental variables generally generate specific gradients that depend upon two main parameters: the

saltwater/freshwater balance and the degree of confinement (Miller et al., 1990; Guelorget and Perthuisot,

1992). Our results on glass eel and elver distribution and spreading around the lagoon arise from this
12
 ACCEPTED MANUSCRIPT
structuring of environmental gradients across the lagoon. Once entered, glass eels remain in the area under

marine influence in the initial phase of recruitment, consistent with the results obtained by other authors (e.g.

White and Knights 1997; Feunteun et al., 2003; Edeline et al., 2007; Trancart et al., 2014). During this

phase, glass eels are still unpigmented, and they undergo a period of adaptation that involves onset of

pigmentation, swim bladder activation and beginning of active swimming, or the so called counter-current

PT
swimming (Trancart et al., 2014). The definitive settlement, when elvers move across nearby areas under

freshwater influence, confirms the high importance of the freshwater appeal for driving movements of

RI
juvenile eels. According to Tosi et al. (1990), Tosi and Sola (1993) and Sola and Tongiorgi (1996), this is

due to specific olfactory cues attracting elvers, which has also been pointed out by Jellyman and Lambert

SC
(2003) and Crivelli et al. (2008). Freshwater appeal causes young eel to move differently in lagoons with

respect to what is observed in estuarine and river systems. This causes a redistribution and settlement of eel

U
along the specific environmental gradients inside the lagoon.
AN
An important parameter to be considered is tide, which is coupled to the moon phase, a relevant factor in
M

controlling glass eel upstream migration in estuaries (Gascuel, 1986; Elie and Rochard, 1994; Tesch, 2003).

The influence of tidal cycles and moonlight on glass eel movement has been well documented (Sorensen and
D

Bianchini 1986, Ciccotti et al., 1995; Jellyman and Lambert, 2003; Crivelli et al., 2008; Zompola et al.,
TE

2008; Sullivan et al., 2009). In our study, even though, glass eels have been observed in respect with new

moon and first quarter moon, model predictions did not reveal a correlation. This factor might therefore not
EP

play a role in driving glass eel movement when already inside the lagoon, contrary to what happens for glass

eels recruiting in an estuarine area or in a channel where they still move by Selective Tidal Transport (STT)
C

(Wippelhauser and McCleave, 1987; for review see Harrison et al., 2014). Therefore, tide will probably
AC

influence glass eel recruitment most likely at the interface sea-lagoon, specifically at the entrance of the

channels connecting the lagoon to the sea. Whereas tidal amplitude along Mediterranean coasts is reduced, it

still induces tidal currents through the sea channels, which allow the exchange of water, dissolved elements

and suspended matter, as well as organisms between the lagoon systems and the open sea. To tidal currents,

water movements driven by winds and by coastal drift will also add, enhancing or reducing the exchange.

Compared to Atlantic estuaries and lagoons, however, Mediterranean coastal lagoons are overall less

affected by the tide (Elliott and Hemingway, 2002). This can explain why we failed to observe a clear
13
 ACCEPTED MANUSCRIPT
relationship between glass eel presence and coupled moon and tide. The modest flows in the channels and

the limited tidal range of most lagoons - especially in the Mediterranean Sea, where tidal currents are too

weak to orientate massively glass eels towards the inlets to lagoons systems - may quantitatively limit glass

eel and other marine fish recruitment. Hence, in coastal lagoons, recruitment of marine fish such as eel is

more affected by the efficiency of the sea connections due to their morphology and management regime and

PT
to its resulting hydro-dynamism (Crivelli et al., 2008).

A factor possibly influencing juvenile eel settlement in lagoons is food availability. Several authors have

RI
suggested that eels do not feed throughout metamorphosis up until their arrival and definitive settlement in

brackish and freshwaters (Tesch, 1977; Lee and Lee, 1989; Lecomte-Finiger, 1992; Kawakami et al., 1999).

SC
The results of this study suggest weak interactions between presence of both zoo-benthic prey and glass eel;

among the potential preys, Chironomids and M. insidiosum was shown to be associated with a high

U
probability of presence of glass eels. These are the only taxa tolerant to low salinities, and their distribution
AN
in the lagoon might be driven by the same environmental gradients that drive glass eels, which could be the

possible explanation of these finding.

M

Contrary to what was expected, the presence of potential preys does not seem to play any role in driving
D

neither elver distribution, nor bootlace eel throughout the lagoon. This may be explained by taking into
TE

account the feeding habits of eel at these stages. Young eels are opportunist feeders consuming benthos.

Temperature and food supply have been widely individuated as environmental factors related to the growth
EP

of eel, because of the temperature role in influencing eel metabolism, mobility and therefore search for food

(Yalçin-Özdilek et al., 2006). On the basis of these observations, our findings on ELV and BTL feeding
C

activity could be explained because of the inhibition due to the low temperature of waters, and for a failure
AC

of synchrony, in terms of seasons and stations, among presence/density of eels and presence/density of

potential preys (Dorner et al., 2009).

On the whole, the present study has provided information on the behavioural and environmental processes

modulating eel recruitment and settlement to coastal lagoons. These habitats are particularly important for

this species, characterized by an overall plasticity of habitat use, which begins at the juvenile stage, as

confirmed by our findings. This plasticity is particularly evident in lagoon environments, where growth rates

14
 ACCEPTED MANUSCRIPT
and habitat use show typical patterns that are different from those observed in rivers and coastal habitats at

more northern latitudes (Capoccioni et al., 2014), which also concerns recruitment and settlement dynamics.

The prevailing habitat use behaviour in Mediterranean transitional waters is residency, whereas nomadic

behaviour is less frequent (Capoccioni et al., 2014). Moreover, with respect to Northern European

environments, high productive coastal lagoons provide a more favourable condition for growth, and therefore

PT
show higher growth rates (Melia et al., 2006; Bevacqua et al., 2012) as phenotypic growth is positively

correlated with latitude and temperature (Vøllestad, 1992). Again, the knowledge of how fish use their

RI
habitat in space and time throughout their life span appears to be fundamental for understanding their

ecology and population dynamics, and to determine habitat properties important for the development of

SC
effective conservation measures (Gillanders, 2003).

For eel, a management framework aimed at the recovery of the global stock is currently ongoing under a

U
specific European Regulation (CE Regulation 1100/2007). It has been suggested that Mediterranean coastal
AN
lagoons might play an important role (Dekker, 2003b; ICES, 2007; Aalto et al., 2015), because they

constitute an important share of the overall continental habitat of the eel. The general decrease in recruitment
M

has been one of the causes of the decline in eel production in Mediterranean coastal lagoons (Aalto et al.,
D

2015). Other causes are the progressively declining water quality, overfishing and the reduced management
TE

efficiency that also have affected overall fish production and hence eel (Perez Ruzafa and Marcos, 2012,

Katselis et al., 2013) in these sensitive habitats. If on the one hand the overall level of recruitment determines
EP

the general order of magnitude of recruitment, it should on the other hand also apply that site-specific factors

play a significant role in determining fine scale distribution patterns of recruitment at an individual
C

catchment level. Our results point to the fact that an optimal management of coastal lagoons, in particular
AC

maintaining the efficiency of tidal channels and preserving freshwater inputs, might prove to be a measure to

sustain juvenile eel recruitment at the local level, thus sustaining local stocks which has also been pointed

out by Crivelli et al., 2008. This, coupled to measures aiming at sustaining and protecting the escapement of

breeders from coastal lagoons in the Mediterranean could contribute to restoring the stock of this unique

species.

15
 ACCEPTED MANUSCRIPT
Acknowledgements

This work was supported by the Ministero per le Politiche Agricole, Alimentari e Forestali, Direzione

Generale della Pesca e dell’Acquacoltura, Italy, grant “Eel assessment in Italy for the year 2012” to Eleonora

Ciccotti. The collaboration between Dipartimento di Biologia Università degli Studi di Roma “Tor Vergata”

and Dipartimento di Scienze Ambientali, Informatica e Statistica Università “Ca’ Foscari” occurred within

PT
the PRIN (Ministero dell'Istruzione, dell'Università e della Ricerca) project 2009 “Integrated Approaches To

The Ecological Status Assessment And Fishery Management InTransitional Water Ecosystems”, grant to

RI
Eleonora Ciccotti and Piero Franzoi. We are grateful to the Parco Nazionale del Circeo and Ufficio

Territoriale per la Biodiversità di Fogliano for their hospitality and assistance, and to Andrea Fusari and

SC
Enrico Scalchi for support during the sampling campaigns. We are grateful to two anonymous reviewers for

their comments that improved the final version of the manuscript.

References
U
AN
Aalto, E., Capoccioni, F., Terradez Mas, J., Schiavina, M., Leone, C., De Leo, G. A. and Ciccotti, E..
M

Quantifying sixty years of declining European eel (Anguilla anguilla L., 1758) fishery yields in

Mediterranean coastal lagoons. ICES Journal of Marine Science. In press.

D

Able, K.W., Neuman, M.J. andWennhage, H., 2005. Ecology of juvenile and adult stages of flatfishes:
TE

distribution and dynamics of habitat associations. In: Gibson, R.N. Ed., Flatfishes: Biology and

Exploitation. Blackwell Science, Oxford, pp. 164–184.

EP

Als, T. D., Hansen, M. M., Maes, G. E., Castonguay, M., Riemann, L., Aarestrup, K. I. M., ... and

Bernatchez, L., 2011. All roads lead to home: panmixia of European eel in the Sargasso Sea. Molecular
C

Ecology, 20 (7), 1333-1346.

AC

Aranburu, A., Diaz, E. and Briand, C., 2015. Glass eel recruitment and exploitation in a southern European

estuary (Oria, Bay of Biscay). ICES Journal of Marine Science: Journal du Conseil, fsv116.

Arribas, C., Fernández-Delgado, C., Oliva-Paterna, F. J. and Drake, P, 2012. Oceanic and local

environmental conditions as forcing mechanisms of the glass eel recruitment to the southernmost

European estuary. Estuarine, Coastal and Shelf Science, 107, 46-57.

August, S.M. and Hicks B.J., 2008.Water temperature and upstream migration of glass eels in New Zealand:

implica- tions of climate change. Environmental Biology of Fishes, 81, 195–205.

16
 ACCEPTED MANUSCRIPT
Austin, M.P., 2002. Case studies of the use of environmental gradients in vegetation and fauna modeling:

theory and practice in Australia and New Zealand. In: Scott, J.M., Heglund, P.J., Morrison, M.L.,

Haufler, J.B., Raphael, M.G., Wall, W.A., Samson, F.B. Eds., Predicting species occurrences: issues of

accuracy and scale. Island Press, Covelo, California, pp. 73–82.

Bardonnet, A. and Riera, P., 2005. Feeding of glass eels Anguilla anguilla in the course of their estuarine

PT
migration: new insights from stable isotope analysis. Estuarine, Coastal and Shelf Science, 631, 201-

209.

RI
Beaulaton L. and Castelnaud G., 2005. The efficiency of selective tidal stream transport in glass eel entering

the gironde (France). Bulletin Francais de la Pêche et de la Protection des Milieux Aquatiques, 5–21.

SC
Bertin, L., 1956. Eels, a biological study. Cleaver-Hume Press, London.

U
Bevacqua, D., Capoccioni, F., Melià, P., Vincenzi, S., Pujolar, J. M., De Leo, G. A. and Ciccotti, E., 2012.
AN
Fishery-induced selection for slow somatic growth in European Eel. PloSone, 75.

Bonhommeau, S., Chassot, E. and Rivot, E., 2008. Fluctuations in European eel (Anguilla anguilla)
M

recruitment resulting from environmental changes in the Sargasso Sea. Fisheries Oceanography, 17 (1),

32-44.
D

Burke, J. S., Kenworthy, W. J. and Wood, L., 2009. Ontogenetic patterns of concentration indicate lagoon
TE

nurseries are essential to common grunts stocks in a Puerto Rican bay. Estuarine, Coastal and Shelf

Science, 81 (4), 533-543.

EP

Burnham, K.P. and Anderson, D. R., 2002. Model selection and multimodel inference: a practical

information-theoretic approach. Springer.

C

Capoccioni, F., Lin, D., Iizuka, Y., Tzeng, W.-N. and Ciccotti, E., 2014. Phenotypic plasticity in habitat use
AC

and growth of the European eel Anguilla anguilla in transitional waters in the Mediterranean area.

Ecology of Freshwater Fish 23, 65–76.

Cataudella, S., Crosetti, D. and Massa, F., 2014. Mediterranean coastal lagoons: sustainable management

and interactions among aquaculture, capture fisheries and environment. Eds. General Fisheries

Commission for the Mediterranean. Studies and Reviews N. 95. Rome, FAO. 2014. 288 pp.

17
 ACCEPTED MANUSCRIPT
Ciccotti, E., Ricci T., Scardi M., Fresi E. and Cataudella S., 1995. Intraseasonal characterization of glass eel

migration in the River Tiber: space and time dynamics. Journal of Fish Biology, 47, 248–255.

Crivelli, A.J., Auphan, N., Chauvelon, P., Sandoz, A., Menella, J.Y. and Poizat, G., 2008. Glass eel

recruitment, Anguilla anguilla L., in a Mediterranean lagoon assessed by a glass eel trap: factors

explaining the catches. Hydrobiologia 602, 79–86.

PT
Day, J.W., 1989. Estuarine ecology. John Wiley & Sons.

Dekker, W. 2003a. Did lack of spawners cause the collapse of the European eel, Anguilla anguilla?

RI
Fisheries Management and Ecology, 10: 365–376.

Dekker, W. 2003b. Status of the European eel stock and fisheries. Pages 237–254 in K. Aida, K.

SC
Tsukamoto, and K. Yamauchi, editors. Eel biology. Springer Verlag, Tokyo.

Doneddu, M. and Trainito, E., 2005. Conchiglie del Mediterraneo. Il Castello. 253 pp.

U
Dörner, H., Skov, C., Berg, S., Schulze, T., Beare, D. J. and Van der Velde, G., 2009. Piscivory and trophic
AN
position of Anguilla anguilla in two lakes: importance of macrozoobenthos density. Journal of fish

biology, 74(9), 2115-2131.

M

E.C., 2007. European Council Regulation EC No 1100/2007 of 18 September 2007 establishing measures
D

for the recovery of the stock of European eel. Official Journal of the European Union L248/17-23 of
TE

22.9.2007.

Eastwood, P.D., Meaden, G.J., Carpentier, A. and Rogers, S.I., 2003. Estimating limits to the spatial extent
EP

and suitability of sole (Solea solea) nursery grounds in the Dover Strait. Journal of Sea Research, 50,

151–165.
C

Edeline, E., 2007. Adaptive phenotypic plasticity of eel diadromy. Marine Ecology Progress Series, 341,
AC

229-232.

Edeline, E., Lambert, P., Rigaud, C. and Elie, P., 2006. Effects of body condition and water temperature on

Anguilla anguilla glass eel migratory behaviour. Journal of Experimental Marine Biology and Ecology,

331, 217–225.

Elie, P. and Rochard E., 1994. Migration des civelles d’anguilles Anguilla anguilla L. dans les estuaires,

modalities du phenomene et caracteristiques des individus. Bulletin Francais de Peche et de

Pisciculture, 335, 81–98.

18
 ACCEPTED MANUSCRIPT
Elith, J. and Graham, C.H., 2009. Do they? How do they? Why do they differ? On finding reasons for

differing performances of species distribution models. Ecography, 32, 66–77.

Elliott, M. and Hemingway, K.L., 2002. Fishes in Estuaries. Blackwell Science.

Fauvel, P., 1923. Polychètes errantes. Faune de France. Paris. 488 pp.

Fauvel, P., 1927. Polychètes sédentaires. Faune de France. Paris. 494 pp.

PT
Feunteun, E., Laffaille, P., Robinet, T., Briand, C., Baisez, A., Olivier, J. M. and Acou, A., 2003. A review

of upstream migration and movements in inland waters by anguillid eels: toward a general theory. In Eel

RI
biology (pp. 191-213). Springer Japan.

Fielding, A. and Bell, J., 1997. A review of methods for the assessment of prediction errors in conservation

SC
presence/absence models. Environmental Conservation 24, 38–49.

Finiger, R., 1976. Contribution à l’étude biologique et écologique des civelles (Anguilla anguilla Linneus,

U
1758) lors de leur pénétration dans un étang Méditerranéen. I. Recrutement et Biométrie au cours d’un
AN
cycle annuel. , 26: 123–144.

Franco, A., Elliott, M., Franzoi, P. and Torricelli, P., 2008. Life strategies of fishes in European estuaries:
M

the functional guild approach. Marine Ecology Progress Series, 354, 219–228.
D

Franco, A., Fiorin, R., Zucchetta, M., Torricelli, P. and Franzoi, P., 2010. Flounder growth and production
TE

as indicators of the nursery value of marsh habitats in a Mediterranean lagoon. Journal of Sea

Research, 64 (4), 457-464.

EP

Freckleton, R. P., 2011. Dealing with collinearity in behavioural and ecological data: model averaging and

the problems of measurement error. Behavioral ecology and socio-biology, 651, 91-101.
C

Gandolfi, G., M. Pesaro and P. Tongiorgi. 1984. Environmental factors affecting the ascent of elvers
AC

Anguilla anguilla (L.), into the Arno River. Oebalia 10: 17–35.

Gascuel, D., 1986. Flow carried and active swimming migration of the glass eel (Anguilla anguilla) in the

tidal area of a small estuary on the French Atlantic coast. Helgolander Meeresunters, 40, 321–326.

Gascuel, D., Feunteun, E. and Fontenelle, G., 1995. Seasonal dynamics of estuarine migration in glass eels

(Anguilla anguilla). Aquatic Living Resources, 8 (02), 123-133.

Gibson, R.N., 1997. Behaviour and the distribution of flatfishes. Journal of Sea Research, 37, 241–256.

19
 ACCEPTED MANUSCRIPT
Gravina, M. F., and Somaschini, A., 1990. Censimento dei policheti dei mari italiani: Capitellidae Grube,

1862. Atti della Societa Toscana di Scienze Naturali, Memorie, Serie B 97, 259-285.

Guelorget O. and Perthuisot J.P., 1992. Paralic ecosystems, biological organisation and functioning. Vie

Milieu, 42, 215–251

Guisan, A. and Zimmermann, N., 2000. Predictive habitat distribution models in ecology. Ecological

PT
Modelling, 135, 147-186.

Gutiérrez-Estrada, J. C. and Pulido-Calvo, I., 2015. Is the Atlantic surface temperature a good proxy for

RI
forecasting the recruitment of European eel in the Guadalquivir estuary? Progress in

Oceanography, 130, 112-124.

SC
Haro, A.J. and Krueger W.H., 1988. Pigmentation, size, and migration of elvers Anguilla rostrata (Le

Sueur) in a coastal Rhode Island stream. Canadian Journal of Zoology, 66, 2528–2533.

U
Harrison, A. J., Walker, A. M., Pinder, A. C., Briand, C., and Aprahamian, M. W., 2014. A review of glass
AN
eel migratory behaviour, sampling techniques and abundance estimates in estuaries: implications for

assessing recruitment, local production and exploitation. Reviews in Fish Biology and Fisheries, 1-17.
M

Hastie, T.J. and Tibshirani, R.J. Eds., 1990. Generalized Additive Models. Chapman and Hall, New York.
D

335 pp.
TE

Hickman, R.A. 1981. Densities and swimbladder development of juvenile American eels Anguilla rostrata

(Le Sueur) as related to energetics of migration. Journal of Fish Biology, 18, 507–517.
EP

Hiemstra, P.H., Pebesma, E.J., Twenhöfel C.J.W. and Heuvelink G.B., 2008. Automatic real-time

interpolation of radiation hazards: prototype and system architecture considerations. International

C

Journal of Spatial Data Infrastructures Research, 3, 58–72.

AC

ICES (International Council for Exploration of the Sea) 1999. Report of the ICES Advisory Committee on

Fisheries Management, 1998. ICES Cooperative Research Report, 229: 393–405.

ICES (International Council for Exploration of the Sea) 2001. Report of the ICES/EIFAC Working Group

on Eels. ICES CM 2001/ACFM:03.

ICES (International Council for Exploration of the Sea) 2002. Report of the ICES/EIFAC Working Group

on Eels. ICES CM 2002/ACFM:03.

20
 ACCEPTED MANUSCRIPT
ICES (International Council for Exploration of the Sea) 2004. Report of the ICES/EIFAC Working Group

on Eels. ICES CM 2004/ACFM:09.

ICES (International Council for Exploration of the Sea) 2006. Report of the ICES/EIFAC Working Group

on Eels. ICES CM 2006/ACFM:16.

ICES (International Council for Exploration of the Sea) 2007. Report of the ICES/EIFAC Working Group

PT
on Eels. ICES CM 2007/ACFM:23.

ICES (International Council for Exploration of the Sea) 2009. Report of the ICES/EIFAC Working Group

RI
on Eels. ICES CM 2009/ACOM: 15

ICES (International Council for Exploration of the Sea) 2011. Report of the ICES/EIFAC Working Group

SC
on Eels.. ICES CM 2011/ACOM:18.

ICES (International Council for Exploration of the Sea) 2012. Report of the ICES/EIFAC Working Group

on Eels. ICES CM 2012/ACOM:18.

U
AN
Jellyman, D. J., and Lambert, P.W., 2003. Factors affecting recruitment of glass eels into the Grey River,

New Zealand. Journal of Fish Biology, 635, 1067-1079.

M

Jessop, B.M., 2003. Annual variability in the effects of water temperature, discharge, and tidal stage on the
D

migration of American eel elvers from estuary to river. American Fisheries Society Symposium, 33, 3–16
TE

Jimenez-Valverde, A. and Lobo, J.M., 2007. Threshold criteria for conversion of probability of species

presence to either-or presence–absence. Acta Oecologica, 31, 361–369.

EP

Katselis, G. N., Moutopoulos, D. K., Dimitriou, E. N. and Koutsikopoulos, C., 2013. Long-term changes of

fisheries landings in enclosed gulf lagoons (Amvrakikos gulf, W Greece): Influences of fishing and other
C

human impacts. Estuarine, Coastal and Shelf Science, 131, 31-40.

AC

Kawakami, Y., Mochioka, N., Kimura, R., Nakazono, A., 1999. Seasonal changes of the RNA/DNA ratio,

size and lipid contents and immigration adaptability of Japanese glass-eels, Anguilla japonica, collected

in Northern Kyushu, Japan. Journal of Experimental Marine Biology and Ecology, 238, 1–19.

Kettle, A.J. and Haines, K., 2006. How does the European eel (Anguilla anguilla) retain its population

structure during its larval migration across the North Atlantic Ocean? Canadian Journal of Fisheries and

Aquatic Sciences, 63,90–106.

21
 ACCEPTED MANUSCRIPT
Knights, B., 2003. A review of the possible impacts of long-term oceanic and climate changes and fishing

mortality on recruitment of anguillid eels of the Northern Hemisphere. Science of the Total

Environment,310, 237–244.

Le Pape, O., Holley, J., Guerault, D., Desaunay, Y., 2003. Quality of coastal and estuarine essential fish

habitats: estimations based on the size of juvenile common sole (Solea solea L.). Estuarine Coastal and

PT
Shelf Science, 58, 793–803.

Lecomte-Finiger, R. and Razouls, C., 1981. Influence des facteurs hydrologiques et météorologiques sur la

RI
migration anadrome des civelles dans le Golfe du Lion. Cahiers du Laboratoir ed’Hydrobiologie de

Montereau 12, 13-16.

SC
Lecomte-Finiger, R., 1992. Growth history and age at recruitment of European glass eels (Anguilla

anguilla) as revealed by otolith microstructure. Marine Biology 114, 205–210.

U
Lee, T.W. and Lee, K.S., 1989. Daily growth increments and lunar pattern in otolith of the eel, Anguilla
AN
japonica, in the freshwater. Bulletin of the Korean Fisheries Society, 22, 36–40.

Luers, D.F., Love, J.W. and Bath-Martin, G., 2011. Settlement and pigmentation of glass eels Anguilla
M

rostrate Lesueur in a coastal lagoon. Environmental Biology of Fishes 901, 19-27.

D

Manzo, C. 2010. Fish assemblages in three Mediterranean coastal lagoons: structure, functioning and
TE

spatial-temporal dynamics. Ph.D. Thesis, Università degli Studi di Roma “Tor Vergata”, unpublished.

Mariani, S., 2001. Can spatial distribution of ichthyofauna describe marine influence on coastal lagoons? A
EP

central Mediterranean case study. Estuarine, Coastal and Shelf Science, 52 (2), 261-267.

Melia, P., Bevacqua, D., Crivelli, A. J., De Leo, G. A., Panfili, J. and Gatto, M., 2006. Age and growth of
C

the European eel Anguilla anguilla in the Camargue lagoons. Journal of Fish Biology, 68, 876–890.
AC

Miller, J.M. Pietrafesa, L.J. Smith, N.P. 1990. Principles of hydraulic management of coastal lagoons for

aquaculture and fisheries. FAO Fisheries Technical Paper. No. 314. Rome, FAO, 88p.

Miller, M. J., Bonhommeau, S., Munk, P., Castonguay, M., Hanel, R and McCleave, J. D., 2014. A century

of research on the larval distributions of the Atlantic eels: a re‐examination of the data. Biological

Reviews.

22
 ACCEPTED MANUSCRIPT
Minello, T. J., Able, K. W., Weinstein, M. P. and Hays, C. G. 2003. Salt marshes as nurseries for nekton:

testing hypotheses on density, growth and survival through meta-analysis. Marine Ecology Progress

Series, 246, 39-59.

Moriarty, C. and Dekker W., 1997. Management of the European Eel Second Report of the EU Concerted

Action AIR A94- 1939. Fisheries Bulletin No. 15, Dublin: The Marine Institute. p. 52.

PT
Pacariz, S., Westerberg, H. and Björk, G., 2014. Climate change and passive transport of European eel

larvae. Ecology of Freshwater Fish, 23 (1), 86-94.

RI
Pérez-Ruzafa, A. and Marcos C., 2012. Fisheries in coastal lagoons: An assumed but poorly researched

aspect of the ecology and functioning of coastal lagoons. Estuarine, Coastal and Shelf Science, 110, 15-

SC
31.

Pihl, L.A., Cattrijsse, I., Codling S., Mathieson, D.S., McLusky and Roberts, C., 2002. Habitat Use by

U
Fishes in Estuaries and Other Brackish Areas. In: Fishes in Estuaries Edited by Michael Elliott, Krystal
AN
Hemingway, Blackwell Publishing Ltd.

Progetto Laghi costieri, 1985. Seminario informativo sui risultati del progetto “Laghi Costieri”.
M

Amministrazione Provinciale di Latina. Università degli Studi “La Sapienza” di Roma.

D

R Development Core Team, 2014. R: A language and environment for statistical computing. R Foundation
TE

for Statistical Computing, Vienna, Austria. http://www.R- project.org.

Robinson, L.M., Elith, J., Hobday, a. J., Pearson, R. G., Kendall, B. E., Possingham, H. P. and Richardson,
EP

A.J., 2011. Pushing the limits in marine species distribution modelling: lessons from the land present

challenges and opportunities. Global Ecology and Biogeography. doi:10.1111/j.1466-8238.2010.00636.x

C

Ruffo S. (ed.). 1982-1888. The Amphipoda of the Mediterranean. Part 1, Part 2, Part 3, Part 4. Mémoires de
AC

l'Institut océanographique, Monaco, 13, 1-959.

Sconfietti, R., 2004. Chiave di riconoscimento dei più comuni peracaridi Crustacea, Peracarida lagunari

italiani. Studi Trentini di Scienze Naturali., 81, 79-89.

Signorini, A., Massini, G., Migliore, G., Tosoni, M., Varrone, C. and Izzo, G., 2008. Sediment

biogeochemical differences in two pristine Mediterranean coastal lagoons (in Italy) characterized by

different phanerogam dominance. A comparative approach. Aquatic Conservation: Marine and

Freshwater Ecosystems, 18 (S1), S27-S44.

23
 ACCEPTED MANUSCRIPT
Silberschneider, V., Pease, B.C. and Booth, D.J. 2001. A novel artificial habitat collection device for

studying resettlement patterns in anguillid glass eels. Journal of Fish Biology, 58, 1359-1370.

Sola, C. and Tongiorgi, P., 1996. The effect of salinity on the chemotaxis of glass eels, Anguilla anguilla, to

organic earthy and green odorants. Environmental Biology of Fishes, 472, 213-218.

Sorensen, P.W. and Bianchini M.L., 1986. Environmental correlates of the freshwater migration of elvers of

PT
the American eel in a Rhode Island brook. Transactions of the American Fisheries Society, 1152, 258-

268.

RI
Stoner, A.W., Manderson, J.P. and Pessutti, J.P., 2001. Spatially explicit analysis of estuarine habitat for

juvenile winter flounders: combining generalized additive models and geographic information systems.

SC
Marine Ecology Progress Series, 213, 253–271.

Stoner, A.W., Spencer, M.L. and Ryer, C.H., 2007. Flatfish-habitat associations in Alaska nursery grounds:

U
use of continuous video records for multi-scale spatial analysis. Journal of Sea Research, 57, 137–150.
AN
Strubberg, A., 1913. The metamorphosis of elvers as influenced by outward conditions. Some experiments.

Meddelander fra Danmarks Fiskeri og Havundersøgelser, Copenhagen: 1-11.

M

Sullivan, M.C., Wuenschel, M.J. and Able, K.W., 2009. Inter- and intra-estuary variability in ingress,
D

condition and settlement of the American eel Anguilla rostrata: implications for estimating and
TE

understanding recruitment. Journal of Fish Biology, 74, 1949–1969.

Tesch, F.W. 2003. The Eel. Blackwell Science.

EP

Tesch, F.W., 1977. The eel: biology and management of Anguillid eels. Chapman and Hall, London.

Tosi, L., Spampanato, A., Sola, C. and Tongiorgi, P., 1990. Relation of water odour, salinity and
C

temperature to ascent of glass‐eels, Anguilla anguilla (L.): a laboratory study. Journal of Fish
AC

Biology, 36 (3), 327-340.

Tosi, L. and Sola C., 1993. Role of geosmin, a typical inland water odor, in guiding glass eel Anguilla

anguilla L migration. Ethology, 95, 177-185.

Trancart, T., Lambert, P., Daverat, F. and Rochard, E., 2014. From selective tidal transport to counter-

current swimming during watershed colonisation: an impossible step for young-of-the-year catadromous

fish? Knowledge and Management of Aquatic Ecosystems, 412, 04.

24
 ACCEPTED MANUSCRIPT
van Ginneken, V., Antonissen, E., Müller, U.K., Booms, R., Eding, E., Verreth, J. and van den Thillart, G.,

2005. Eel migration to the Sargasso: remarkably high swimming efficiency and low energy costs.

Journal of Experimental Biology, 208, 1329–1335.

Vaz, S., Martin, C.S., Eastwood, P.D., Ernande, B., Carpentier, A., Meaden, G.J. and Coppin, F., 2008.

Modelling species distributions using regression quantiles. Journal of Applied Ecology,45, 204–217.

PT
Vollestad, L.A., 1992. Geographic variation in age and length at metamorphosis of maturing European eel:

environmental effects and phenotypic plasticity. Journal of Animal Ecology, 61, 41–48.

RI
Walsh, C. T. and Pease, B. C., 2002. The use of clove oil as an anaesthetic for the longfinned eel, Anguilla

SC
reinhardtii (Steindachner). Aquaculture Research, 33, 627–635.

White, E. M. and Knights, B., 1997. Dynamics of upstream migration of the European eel, Anguilla

U
anguilla (L.), in the Rivers Severn and Avon, England, with special reference to the effects of man‐made
AN
barriers. Fisheries Management and Ecology, 4 (4), 311-324.

Wippelhauser, G. S. and McCleave, J. D., 1987. Precision of behavior of migrating juvenile American eels
M

(Anguilla rostrata) utilizing selective tidal stream transport. Journal du Conseil: ICES Journal of Marine
D

Science, 44 (1), 80-89.

TE

Wood, S.N. 2006. Generalized Additive Models: an Introduction with R. Chapman and Hall–CRC, Boca

Raton. 392 pp.

EP

Yalçin, Ş., Gümüş, A. and Dekker, W., 2006. Growth of European eel in a Turkish river at the south-eastern

limit of its distribution. Electronic Journal of Ichthyology, 2, 55-64.

C

Zompola, S., Katselis, G., Koutsikopoulos, C. and Cladas, Y., 2008. Temporal patterns of glass eel
AC

migration Anguilla anguilla L. 1758 in relation to environmental factors in the Western Greek inland

waters. Estuarine, Coastal and Shelf Science, 80, 330-338.

Zucchetta, M., Franco, A., Torricelli, P. and Franzoi, P., 2010. Habitat distribution model for European

flounder juveniles in the Venice lagoon. Journal of Sea Research, 64 (1), 133-144.

Zurell, D., Jeltsch, F., Dormann, C.F. and Schröder, B., 2009. Static species distribution models in

dynamically changing systems: how good can predictions really be? Ecography 32, 733–744.

25
 ACCEPTED MANUSCRIPT
Tables

Table 1. Results of the three sampling campaigns.

Days of presence during the sampling campaign, abundance ranges (total number and mean value of individuals) and

range of length are shown for each sampling campaign (December, February and April) and for each stage (GLE =

glass eel; ELV= elver; BTL= bootlace), specifying where (station number) and how (gears, T= traps; GEfy= glass eel

fyke nets; Efy= eel fyke nets) samples were taken.

PT
GLE ELV BTL
Days of presence 7 2 1
Total number (mean) 28 (0.6) 3 (0.1) 1

RI
December Range of length 46-70 mm 58-71 mm 15 cm
Stations 1,2,3,7 3,4 5
Gears T T, GEfy GEfy, Efy

SC
Days of presence 7 4 3
Total number (mean) 311 (6.3) 16 (0.3) 3 (0.1)
February Range of length 50-65 mm 57-68 mm 13-15 cm
Stations 1,3,7 2,3,4,5 4,5,6

U
Gears T, GEfy GEfy GEfy, Efy
Days of presence 1 2 6
AN
Total number (mean) 1 2 (0.0) 22 (0.4)
April Range of length 56 mm 58-65 mm 12-15 cm
Stations 7 4,7 1,2,3,5,6,7
Gears T, GEfy T T, GEfy, Efy
M
D
TE
C EP
AC

26
 ACCEPTED MANUSCRIPT

Table 2. Environmental variables.

Mean values and range of variation of environmental variables for each month of sampling and each station.

Station Temperature (°C) Salinity (PSU) Dissolved Oxygen (mg l-1) Chlorophyll-a (µg l-1)

PT
1 7.7 (3.8 - 10.9) 27.9 (24.6 - 29.8) 8.6 (7.8 - 9.6) 0.4 (0.0 - 0.9)
2 8.2 (6.1 - 10.0) 26.2 (22.6 - 28.0) 8.5 (8.0 - 9.4) 0.9 (0.0 - 1.7)

RI
3 7.2 (4.7 - 10.1) 26.1 (19.5 - 29.2) 8.8 (8.1 - 10.2) 0.9 (0.0 - 1.6)
December 4 7.0 (4.6 - 10.1) 27.5 (23.9 - 29.1) 8.8 (8.2 - 10.1) 1.0 (0.4 - 1.9)

SC
5 6.9 (5.5 - 9.2) 28.1 (26.1 - 29.0) 8.9 (8.5 - 9.9) 0.8 (0.5 - 1.1)
6 7.1 (5.3 - 9.4) 27.7 (24.6 - 29.1) 9.1 (8.5 - 10.3) 0.5 (0.0 - 1.1)
7 7.8 (4.5 - 10.9) 27.5 (21.2 - 30.0) 8.9 (7.9 - 10.6) 0.7 (0.1 - 1.5)

U
AN
1 8.7 (8.2 - 9.4) 22.9 (20.6 - 23.9) 9.1 (8.1 - 11.2) 0.2 (0.0 - 0.4)
2 9.2 (8.6 - 9.7) 21.2 (14.5 - 23.7) 9.6 (8.2 - 13.7) 0.8 (0.0 - 1.8)

M
3 9.1 (8.3 - 9.7) 22.6 (20.5 - 23.7) 9.6 (7.9 - 12.9) 0.8 (0.0 - 3.1)
February 4 9.0 (8.7 - 9.3) 22.0 (19.8 - 24.2) 9.3 (7.8 - 10.5) 0.3 (0.0 - 0.6)
5 9.3 (8.3 - 10.0) 21.8 (20.1 - 23.5) 9.5 (7.8 - 11.1) 1.0 (0.1 - 2.2)

D
6 9.4 (9.1 - 9.9) 21.4 (20.3 - 21.8) 11.8 (9.7 - 13.2) 1.5 (0.6 - 2.6)

TE
7 8.7 (8.2 - 9.3) 23.2 (21.3 - 24.1) 9.7 (7.9 - 12.8) 0.4 (0.0 - 1.0)

1 18.4 (16.8 - 20.1) 25.0 (22.5 - 28.3) 7.5 (7.0 - 8.1) 0.1 (0.0 - 0.6)
EP
2 18.9 (17.5 - 20.6) 24.6 (24.1 - 25.2) 6.8 (5.9 - 7.2) 0.2 (0.0 - 1.2)
3 18.9 (17.4 - 20.3) 23.7 (20.7 - 25.0) 7.4 (6.8 - 7.8) 0.1 (0.0 - 0.2)
C

April 4 18.6 (17.3 - 20.6) 24.1 (22.8 - 24.8) 7.2 (6.7 - 7.6) 0.1 (0.0 - 0.8)
AC

5 19.2 (17.9 - 20.8) 23.4 (22.3 - 24.2) 7.2 (6.6 - 7.6) 0.1 (0.0 - 0.4)
6 19.0 (17.8 - 20.9) 24.2 (23.5 - 24.6) 6.5 (5.7 - 8.3) 1.5 (1.0 - 3.0)
7 18.2 (16.6 - 19.2) 25.2 (24.6 - 25.6) 7.8 (7.5 - 8.2) 0.2 (0.0 - 0.4)

27
 ACCEPTED MANUSCRIPT
Table 3. AICc for all model combinations.

AICc for the combination of the best set of models for each stage is given by the combination of models reported, for

which the sum of the AICc weights is equal to 95%. AICc: Akaike Information Criterion corrected; T: Temperature;

Moon: tidal phase; ∆ inlet: distance from the tidal channel; ∆ FW: distance from the fresh water supply; DO: Oxygen;

Sal: Salinity; Chi: Chironomidae larvae; Min: Monocorophium insidiosum; Gae: Gammarus aequicauda; Mgr:

Microdeutopus gryllotalpa; Nru: Nereiphylla rubiginosa; Pdu: Platynereis dumerilii.

PT
Glass eel (GLE) Elver (ELV) Bootlace (BTL)

RI
Component Model AICc Weight AICc Weight AICc Weight

T 66.28 0.02 73.45 0.02

SC
T+Moon 73.45 0.02

T+DO 73.45 0.02

U
T+Moon+DO 73.45 0.02
AN
Sal 66.23 0.02

∆ FW+∆ inlet 61.51 0.19

M

PSU+DO+∆ FW+∆ inlet 58.78 0.73

T+Moon+Sal+∆ inlet 89.15 0.01 70.93 0.08

D

T+Moon+Sal +∆ FW 69.9 0.14

TE

T+Moon+Sal +∆ inlet+DO 87.81 0.03 70.93 0.08

T+Moon+Sal +∆ FW+DO 69.9 0.14

T+Moon+Chi+Min+Mgr+Gae+Pdu+Nru+DO 85.32 0.1 71.46 0.06

EP

T+Moon+Chi 73.45 0.02

T+Moon+Min+Mgr+Gae 71.46 0.06

C

T+Moon+Pdu+Nru 73.45 0.02

AC

PSU+DO+∆ FW+Chi+Min+Mgr+Gae+Pdu+Nru 90.16 0.01

T+Moon+Sal+DO+∆ inlet+Chi+Min+Mgr+Gae+Pdu+Nru 81.79 0.58 70.22 0.12

T+Moon+Sal+DO+∆ FW+Chi+Min+Mgr+Gae+Pdu+Nru 83.51 0.25 69.87 0.14

28
 ACCEPTED MANUSCRIPT
Figure captions

Figure 1. Study area and sampling stations. Asterisks indicate the outlet of freshwater inputs from channels, now closed

in order to avoid excessive nutrient inputs to the lagoon (see text). At present, occasional freshwater inputs to the lagoon

can occur only from occasional overflows from the perimeter channels running along the inner shore of the lagoon (one

behind station 5, one behind station 2), or at incoming tides near the mouth of river Martino.

PT
Figure 2. Abundances (n) of the three juvenile eel stages during the three sampling campaigns at all sampling stations

(see Fig. 1 for details).

RI
Figure 3. Relationships between predictor variables and probability of presence for glass eel (dotted line), elver (dashed

SC
line) and bootlace (continuous line). Grey bands are 95% point wise confidence intervals.

Figure 4. Model’s prediction of the probability of presence of GLE (a), ELV (b) and BTL (c) in the whole basin. The

U
maps show the monthly evolution of the eel estimated probability of presence during the sampling campaigns carried
AN
out in December, February, March and April.
M
D
TE
C EP
AC

29
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC